Suggestions/requests for those posting Google's AI Overviews

[sl8]

Last edited Thu Aug 21, 2025, 01:03 PM - Edit history (1)

1. Include the prompt you used to get the overview. Very often, changing the prompt very slightly will get you a response nearly completely opposite to that generated by the first prompt

2. Before posting the overview, check Google's source links; they're ften from pretty worthless sources (e.g., social media posts from unknown posters) and sometimes the the sources say things that are very inconsistent with Google's interpretation. I saw a post here the other day that used an AI overview to show that a media source was reputable, but upon checking the source links, all but one originated with the media source in question (unsurprisingly, the media source had nothing but nice things to say about themselves).

3. Reconsider posting a Google AI Overview, at all.

On edit:
I should have started the list with a request to identify the AI material as AI material.

[cachukis]

1. Be mindful, no court has yet to accept the veracity

of an AI analysis.

[ratchiweenie]

2. Yes please, identify AI material as AI so we know to ignore it.

[highplainsdem]

3. Just go with suggestion 3, and make it stronger: Don't post Google AI Overviews at all.

Using, let alone quoting, one of those Overviews communicates certain messages more clearly than whatever the Overview itself says on any subject.

Your use of an AI Overview says:

1) I'm too lazy to do some real research and put it into my own words as a summary, or refer to actual sources.

2) I don't care that AI Overviews are often wrong.

3) I don't care that AI Overviews steal traffic from websites and are INTENDED to.

Nothing that anyone can post from an AI Overview - UNLESS they're posting to warn people how wrong AI Overviews can be - can outweigh what using them says about the user.

Just as using AI to cheat in any way - whether it's cheating in school or having a mindless plagiarism machine "come up with ideas" - says much more about the AI user than it does about anything churned out by the AI.